PPC Agency, Freelancer or In-House? A UK Cost Guide
PPC company, freelance PPC consultant or an in-house hire? Compare true UK costs, cover and skills, with a decision tree and an ownership checklist.
Generative engine optimization (GEO) explained: how to get your brand cited in ChatGPT, Perplexity and Google AI Mode, and how to measure it in 2026.

For twenty years, getting found online meant ranking on Google. Now a growing share of research happens inside answer engines: ChatGPT, Perplexity, Claude, Gemini and Google’s own AI features. In July 2026 Google said AI Mode had passed 1 billion monthly active users and that it had merged AI Overviews and AI Mode into one Search experience.
Generative Engine Optimization (GEO) is the practice of structuring your content so that large language models cite, quote and recommend your brand when people ask buying-intent questions. It sits on top of SEO and doesn’t replace it.
If competitors get named in AI answers and you don’t, you lose more than clicks. You lose the recommendation itself.
The term comes from a 2023 research paper by researchers at Princeton and IIT Delhi, GEO: Generative Engine Optimization (Aggarwal et al., published at KDD 2024). GEO has three goals:
Classic SEO targets one ranked list. GEO targets a generated answer that may or may not include your name, so small structural choices carry weight.
The GEO paper tested content changes on a benchmark of 10,000 queries. Adding quotations from credible sources, adding statistics and citing sources raised a page’s visibility in generated answers the most, by up to about 40% on the authors’ measures (Aggarwal et al.). Keyword stuffing did worse than leaving the page alone.
Google’s guidance points the same way. Its documentation on AI features and your website says there are no extra requirements to appear in AI Overviews or AI Mode. The same SEO basics apply: crawlable pages, helpful content and a good page experience.
In practice, content that gets cited tends to:
If a paragraph can’t stand alone as an answer, a model is unlikely to quote it as one.
Fix the plumbing before you write. AI crawlers behave much like classic search bots.
Allow the right crawlers. Most AI companies run separate bots for search and for training, and you can treat them differently in robots.txt:
OAI-SearchBot surfaces pages in ChatGPT search; GPTBot collects training data (OpenAI crawler docs).Claude-SearchBot for search, ClaudeBot for training, Claude-User for user-requested fetches (Anthropic’s crawler policy).PerplexityBot for search results (Perplexity bot docs).Googlebot. The Google-Extended token only controls use of your content for Gemini training and grounding outside Search, so blocking it doesn’t remove you from AI Overviews.If you want to be cited, don’t block the search bots. Blocking only the training bots is a reasonable choice.
Serve HTML. Ship server-rendered HTML for the content you want quoted. Content that only appears after JavaScript runs is harder for several of these bots to read.
Keep your sitemap clean. A current sitemap.xml helps every crawler.
Treat llms.txt as optional. Some sites publish an llms.txt file that summarizes the site for AI tools. Google says you don’t need “AI text files” to appear in its AI features, and Google’s John Mueller said in June 2025 that no AI system uses llms.txt. It’s cheap to add, but don’t expect results from it.
Use structured data for context. JSON-LD such as Organization, Article and Product helps engines understand your pages. Google says no special schema is needed for its AI features, and it stopped showing FAQ rich results in May 2026, so don’t add FAQPage markup for rich results.
How you write matters more than which tool you use.
Lead with a definition. The first sentence of a piece should answer “what is this?” on its own. That sentence is the one most likely to be lifted.
Use the question as the heading. “What is server-side tracking?” matches how people phrase prompts better than “Server-side tracking explained.”
Write extractable lists. Numbered steps, named frameworks and short comparison tables are easy for a model to pull into an answer.
Cite sources and add numbers. The GEO paper found statistics, quotations and source citations among the strongest levers. Original data from your own work is even better, as long as it’s real and you can show how you got it.
Name your methods. A named method, such as a “five-step CAPI audit,” gives the model a phrase to attach to your brand.
Classic SEO rewards coverage of related keywords. GEO rewards depth on a defined topic. If you want AI engines to cite you on Google Ads for e-commerce, you need a body of serious, interlinked articles on that topic on one domain, not one post and a service page.
Thin content fails to rank, and engines have little reason to cite it.
Yes. Models learn from the public web, and AI search tools retrieve from it. Mentions in newsletters, podcasts, Reddit threads, niche forums and industry roundups shape whether engines connect your brand with a topic. An unlinked mention on a respected site still helps, because the association is in the text.
If your PR goal is “do they link to us?”, change it to “do they name us?”
You can’t measure GEO the way you measure SEO, but the tools are improving.
chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com. Getting this right depends on clean conversion tracking; our post on why your analytics data goes missing covers the gaps.Pick the five questions your highest-intent prospects ask before they buy. Write one thorough article per question using the patterns above, with sources for every number. Check your robots.txt lets the search bots in. Then prompt the engines and log what changes.
Answer engines send people to pages, so the page still has to convert. Treat cited pages like any other entry point and make sure they lead somewhere useful, such as a landing page built to convert. For SaaS companies, whose buyers often research tools through AI assistants, this is worth starting now.
SEO aims to rank pages in a list of search results. GEO aims to get your content quoted or cited inside an AI-generated answer. GEO builds on SEO: if search engines can’t crawl and trust your pages, AI engines that rely on them won’t cite you either.
No. Google says there are no extra requirements or special schema for its AI features. Pages need to be indexed and eligible for a snippet in normal Search.
Block training crawlers such as GPTBot or ClaudeBot if you don’t want your content used for model training. Leave search crawlers such as OAI-SearchBot, Claude-SearchBot and PerplexityBot allowed if you want to be cited in their answers.
No major engine has confirmed that it uses llms.txt, and Google says you don’t need it for its AI features. It does no harm, but it’s not a ranking or citation lever.
There’s no official ChatGPT report for site owners. Run your key questions in ChatGPT on a schedule and log the answers, and track referral sessions from chatgpt.com in your analytics.
We use cookies for analytics and advertising. You can opt out of the sale or sharing of your data at any time.Cookie Policy
Your browser's Global Privacy Control signal is on, so we don't share your data for advertising.